The research proposed here focuses on fundamental mechanisms employed by the brain for the analysis of communication sounds. Humans, and a diversity of other species, use low frequency sounds for communication, and research into the neural mechanisms that extract information encoded by the primary afferent spike trains is fundamental to the development of therapeutic measures for the hearing impaired. The fish, Pollimyrus, is used as a model for examining the neurobiology of auditory temporal processing because the animal's ear is relatively simple and is well adapted for time coding of signals, the fish uses well characterized communication sounds, and hearing can be studied with a combination of approaches including single neuron electrophysiology, anatomy and behavior. The experiments are designed to test hypotheses about the brain mechanisms that produce neurons that are sensitive to particular temporal features of communication sounds. Neurons in the auditory nerve produce trains of spikes that are precisely entrained to the time structure of sounds, but are not feature selective. In contrast, neurons in the midbrain are feature selective and must reflect the output of neural processing networks that produce selectivity from entrained input. A series of physiological and behavioral experiments will be carried out to evaluate five specific hypotheses about the nature of these neural computations: (1) Interval-selectivity is based on the time-intervals between spikes in the afferent input; (2) The mechanism producing selectivity depends upon inhibition and inhibitory rebound; (3) GABA-ergic neurons in the lemniscal nucleus IR.N provide inhibitory input to the selective neurons; (4) The system, as assessed behaviorally, is sensitive to small differences in the periods of naturalistic sounds; (5) Behavioral discrimination depends upon temporal features (i.e. phase spectrum) of periodic stimuli.